Collaborative filtering confusing numbers

Hi all,
I’ve ran the collaborative filtering example, but instead of the Movie Lens Data set, i used my own data set.
My data set has been taken from an online shop orders history Database with same columns:UserId,ItemId,Amount (instead of rating between [0,5], i have [0,1]).
after learning and fitting the model, i Get huge numbers in the Train Loss and the Valid Loss.
Attached print screen from the notebokk.
Can anyone direct me what’s wrong with my work?

Thanks in advance,